There is no coupon and no dividend here. The return is realised only at an acquisition or a listing, most positions return little or nothing, and the entire portfolio's economics usually rest on one or two outcomes out of many.
12 min read · Level 3
The number I would want a new angel to sit with is not 3.55x, it is 70.42% - the share of a whole portfolio's proceeds that a single position out of twenty produced. Most explanations of this asset class lead with the multiple because the multiple is the flattering number, and they treat the concentration as a footnote about variance. I think that gets the emphasis backwards. The concentration is not noise around a real underlying average; it is the mechanism, and a portfolio that has not produced something close to this shape has usually not been given the chance to, either because too few positions were made for the law of large numbers to have any room to operate, or because the follow-on decisions on the positions that started working were made too cautiously to let them compound. What follows from taking the concentration seriously is a discipline that looks almost the opposite of ordinary portfolio management: you are not trying to avoid your worst outcomes, because eleven zeros out of twenty is the expected shape of a well-constructed portfolio in this class, not a sign that something went wrong. You are trying to make sure that when one position starts working, nothing - not a missed pro-rata, not an underfunded reserve, not an eagerness to lock in a partial gain - stops you from being fully exposed to it. That is a genuinely unusual instruction to give an investor, and it is the correct one here.
There is no coupon here, no rent, and in the ordinary case no dividend - a startup that can afford to distribute cash to shareholders is, by the definition most investors in this class use, no longer behaving like the kind of company that justified the risk. The entire return, for every position, arrives in one of two ways: the company is acquired, or it lists its shares publicly. Investors call this moment a liquidity event, and until it happens the position is a number on a cap table, not money.
That structural fact interacts with a second one to produce the defining shape of this asset class. Take a portfolio of 20 positions of 25,000 each, 500,000 invested in total, with outcomes that are illustrative rather than a claim about any real fund's actual record - but the shape is the one this class produces reliably:
Outcome Count Multiple Proceeds Total loss 11 0x 0 Roughly flat 5 1x 125,000 Modest win 2 3x 150,000 Strong win 1 10x 250,000 Standout win 1 50x 1,250,000 ----------------------------------------------- Total 20 1,775,000
Invested 500,000 Gross multiple on invested capital 3.55x
Read the count column before the proceeds column. Eleven of twenty positions - 55.0% - returned nothing at all, and a further five returned roughly what was put in, meaning 16 of 20, or 80.0%, produced no real gain once years of illiquidity are accounted for. The entire portfolio's 3.55x came from three positions out of twenty.
Isolate the single standout. Remove its 1,250,000 from the total and the remaining nineteen positions returned 525,000 against 500,000 invested - a bare 1.05x, effectively flat before any consideration of the years the money was locked up. Remove both the 10x and the 50x outcomes and the remaining eighteen positions returned 275,000 against 500,000 - a 0.55x, a real loss of 45.0%. One investment, one in twenty by count, produced 1,250,000 of the portfolio's 1,775,000 in proceeds: 70.42% of everything the portfolio ever returned. The top two together produced 84.51% of it.
This is not a portfolio that happened to get unlucky on eighteen positions and lucky on two. It is what this asset class does by construction, and it is the reason every later lesson in this subject - pricing, evaluation, sizing - has to be read against a background in which the median outcome, by count, is zero.
Because the size of a return and the speed at which it arrives are two separate questions, and the second one is where this class differs most sharply from anything with a coupon.
Convert the multiples above into annualised terms over a plausible eight-year holding period, since a multiple without a timeframe understates how demanding the standout outcome actually has to be. A 50x return over eight years requires an internal rate of return of 50^(1/8) - 1 = 63.09% a year, compounded, for eight consecutive years. A 3x return over the same eight years requires only 3^(1/8) - 1 = 14.72% a year - a figure that would be a respectable outcome in public equities, and here it is one of the disappointing results in the table. A 1x return over eight years is not neutral; it is a real loss once the years the capital could have earned something else are counted, which is the same point volatility-vs-permanent-loss makes about a public holding that goes nowhere, made sharper here because nothing was paid out along the way to compensate for the wait.
The follow-on mechanism changes the arithmetic again, and it explains why the winners in a real portfolio are rarely accidents of the initial cheque alone. Pro-rata rights - a contractual right, not automatic, to invest further in later rounds to maintain a given ownership percentage - exist precisely because dilution (lesson four's subject) would otherwise erode a winning position's ownership at the exact moment its value is compounding fastest. An investor who cannot or does not exercise pro-rata on their best-performing company gives up a share of the very outcome that is paying for the other nineteen.
And the total-loss line deserves a plainer treatment than a percentage in a table. Unlike a public holding that falls and can, in principle, recover, most of the eleven zeros in this portfolio are not paper losses waiting for a rebound - the company ceased operating, and the position is permanently gone. There is no volatility to sit through and no price to watch; the number simply becomes zero and stays there. That is a different psychological and financial event from a 90% drawdown in a liquid asset, and treating it as the same kind of loss understates how final it usually is.
This is educational information, not investment, securities or financial advice. Rules differ by country and change — get advice from a qualified professional in your jurisdiction before acting.
Educational, not personalised financial advice. Nothing in the Investor Academy is a recommendation to buy, sell or hold anything — what is right for you depends on your own circumstances and the rules where you live.
Sources: U.S. SEC - Investor.gov; FINRA - Investor Insights
Two angels, each having made the same twenty investments in the table above, differ only in what they did with their best position.
Angel one exercised pro-rata on the standout winner at every subsequent round, maintaining their original percentage rather than letting later rounds dilute it. Their share of the eventual 1,250,000 proceeds reflects the ownership they started with. Their total portfolio proceeds: the full 1,775,000 shown above, a 3.55x gross multiple on the 500,000 committed - though note this ignores the additional capital pro-rata itself required, which a full accounting would net against the multiple.
Angel two had the identical initial position in the same company but could not fund the follow-on rounds, and their ownership was diluted by roughly 20% at each of three subsequent financings - the mechanism lesson four covers in full. Their stake in the standout winner, and only that position, ends up at 0.8^3 = 51.2% of what it would otherwise have been. Their proceeds from that one position: 1,250,000 x 0.512 = 640,000 rather than 1,250,000, a shortfall of 610,000.
Angel one, total portfolio proceeds 1,775,000 Angel two, total portfolio proceeds 1,165,000 (1,775,000 - 610,000) Angel two's gross multiple 2.33x (1,165,000 / 500,000)
Both angels picked the identical twenty companies. Both were right about the one that mattered. The entire difference between a 3.55x and a 2.33x outcome was a follow-on decision made years after the first cheque, on a company that was already known to be working - which is precisely why reserving capital for pro-rata, covered again in lesson six's sizing discussion, is not a minor administrative detail in this asset class. It is frequently the difference between capturing the return the whole portfolio was built to produce and giving away a large share of it to dilution nobody was forced to accept.
Judging a single position by its multiple without asking how long the capital was locked up to get there. A 3x that took eight years to arrive is a 14.72% annual return; the same 3x in two years is a 44.22% annual return. The multiple alone cannot distinguish them.
Treating a flat or slightly positive outcome as a success. A position returning 1x after eight years of complete illiquidity is a real loss once the opportunity cost of eight years' capital is counted - the same reasoning volatility-vs-permanent-loss applies to a public holding that goes nowhere, sharpened here by the total absence of any income along the way.
Expecting the portfolio's average position to resemble the portfolio's average return. In the illustrative table, the median outcome by count is zero, and the mean is pulled entirely upward by one or two positions. Anyone sizing a single investment as though it were likely to be "typical" of the 3.55x headline has mistaken the mean for the median in the one asset class where the gap between them is largest.
Declining to exercise pro-rata on a winning position for reasons of convenience. The follow-on decision in the worked example cost 610,000 of proceeds on a portfolio that had already identified its winner - the hardest part of the decision, picking the company, had already been made correctly, and the value was given away afterward through inaction rather than through a wrong initial judgement.
Waiting for a total loss to "come back". Most of the eleven zeros in this class do not recover, because the company has usually stopped operating rather than merely fallen in price. There is no position to hold through; there is only a number that stays at zero.
A multiple with no holding period attached tells you almost nothing about how demanding the outcome was. Over eight years, 50x requires 63.09% a year compounded and 3x requires only 14.72% - work out roughly how long a position actually took before judging it against another.
Write every position's outcome, count how many landed at zero and at roughly 1x, and compute what the portfolio would have returned with its best one or two positions removed. On the numbers here that removes 3.55x down to 0.55x - the single most informative calculation an angel investor can run on their own results.
By the time a company is obviously working, every other investor wants in on the same terms, and reserving the capital to defend your position has to be decided in advance rather than negotiated under pressure at the moment it matters most.
Unless there is a specific, named reason to expect a restructuring or a sale of remaining assets, a company that has stopped operating is not going to produce a recovery. Recording it as a permanent loss of capital rather than an open position is the more honest bookkeeping, and it is the correct one.
When judging a single position's outcome, when deciding whether to commit further capital to a follow-on round, and whenever a portfolio's headline multiple is being quoted without the distribution of individual outcomes behind it.
Do not use the illustrative 20-position table as a claim about what any real portfolio should expect to return - it exists to show the shape of the distribution, not a forecast. And do not use the pro-rata discussion to conclude that every winning position must be defended at any cost; lesson six covers when concentrating further into one company stops being the correct answer.
A frequently observed shape in angel portfolios is that the position which eventually produces most of the return was, by the time of the outcome, owned at a smaller percentage than the investor's original cheque implied - not because the investor sold anything, but because they did not defend their position through the rounds in between. Each individual dilution event, taken alone, looks minor: a percentage point or two at a single financing. Compounded across three or four rounds over several years, on the one position that turned out to matter, the cumulative effect is large enough to materially change the portfolio's overall return, and it is discovered, if at all, only when the final outcome is being tallied - by which point nothing about it can be changed.
The follow-on decision on a working position is not a footnote to the initial investment decision. On the numbers a real portfolio produces, it is frequently worth more than the initial decision was.
Create a free account to read the rest and apply it to your own investing.